Update README.md
Browse files
README.md
CHANGED
|
@@ -1,10 +1,89 @@
|
|
| 1 |
-
|
| 2 |
-
|
| 3 |
-
|
| 4 |
-
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
---
|
| 9 |
|
| 10 |
-
|
|
|
|
| 1 |
+
# Overworld
|
| 2 |
+
|
| 3 |
+
## The Local World Modeling Company
|
| 4 |
+
|
| 5 |
+
**Simulate Anything on Anything** - Open science world models that run on consumer hardware. From mobile phones to gaming PCs, experience state-of-the-art AI models at blazing speeds.
|
| 6 |
+
|
| 7 |
+
[Join Our Discord](https://discord.gg/dX4HW9Pt7Z) | [Visit Our Website](https://wayfarerlabs.ai/) | [Try OWL Eval](https://eval.wayfarerlabs.ai)
|
| 8 |
+
|
| 9 |
+
## Our Mission
|
| 10 |
+
|
| 11 |
+
We're making world models accessible to everyone. Through open science and community collaboration, we're building AI models that run locally on your hardware - no cloud required. Our models achieve hundreds of FPS on RTX 5090s and 60 FPS on NVIDIA Jetson for robotics applications.
|
| 12 |
+
|
| 13 |
+
## Key Technologies
|
| 14 |
+
|
| 15 |
+
### 🚀 Real-time World Generation
|
| 16 |
+
- **60 FPS** on the latest consumer hardware
|
| 17 |
+
- No cloud computing needed. Run our best models on your current PC or laptop
|
| 18 |
+
|
| 19 |
+
### ⚡ Model Distillation
|
| 20 |
+
- Production-ready optimization techniques
|
| 21 |
+
|
| 22 |
+
### 🧪 Human Evaluation Platform
|
| 23 |
+
- [OWL Eval](https://eval.wayfarerlabs.ai) - An open-source evaluation platform for AI-generated videos
|
| 24 |
+
- Reveals critical model failures that automated metrics miss
|
| 25 |
+
- Understand human perception of AI content
|
| 26 |
+
|
| 27 |
+
## Major Research Projects
|
| 28 |
+
|
| 29 |
+
### 3D Consistency
|
| 30 |
+
Depth map integration and point cloud conditioning for spatially coherent world generation.
|
| 31 |
+
|
| 32 |
+
### Long-term Spatiotemporal
|
| 33 |
+
Solving temporal flickering through frame-pair training and optical flow conditioning.
|
| 34 |
+
|
| 35 |
+
### Novel Architectures
|
| 36 |
+
Diffusion transformers for decoding, hierarchical VAE-GANs, and self-forcing for KV cache optimization.
|
| 37 |
+
|
| 38 |
+
### Embedded Devices
|
| 39 |
+
Models intended to run anywhere from Steam Deck to iPhone (soon), achieving real-time performance without cloud infrastructure.
|
| 40 |
+
|
| 41 |
+
## Active Bounties
|
| 42 |
+
|
| 43 |
+
We offer paid bounties for contributions to our open-source projects. Check our [GitHub Bounty Board](https://github.com/orgs/Wayfarer-Labs/projects/3) for current opportunities.
|
| 44 |
+
|
| 45 |
+
### How to Contribute
|
| 46 |
+
|
| 47 |
+
1. Browse active bounties on our [bounty board](https://github.com/orgs/Wayfarer-Labs/projects/3)
|
| 48 |
+
2. Join our [Discord](https://discord.gg/dX4HW9Pt7Z) (1000+ members) or our model-specific [Discord](https://discord.gg/jT93EuasMA) (*Ask for a role!)
|
| 49 |
+
3. Contact @louis or @spencer on Discord to discuss or claim a bounty
|
| 50 |
+
4. Submit your PR and get paid upon merge
|
| 51 |
+
|
| 52 |
+
## Open Source Projects
|
| 53 |
+
|
| 54 |
+
- **World Models** - State-of-the-art models for real-time generation
|
| 55 |
+
- **OWL Eval** - Human evaluation platform for AI-generated content
|
| 56 |
+
- **Diffusion Decoders** - Optimized decoders with depth pruning
|
| 57 |
+
- **Training Pipelines** - Open frameworks for model training
|
| 58 |
+
|
| 59 |
+
## Recent Publications
|
| 60 |
+
|
| 61 |
+
- [Why Human Evaluation is the Missing Piece in World Model Development](https://wayfarerlabs.ai/blog/why-human-evaluation-is-the-missing-piece-in-world-model-development)
|
| 62 |
+
- [Optimizing Diffusion Decoders with Depth Pruning](https://wayfarerlabs.ai/blog/diffdec-ode-regression)
|
| 63 |
+
- [Generation vs. Reconstruction: Striking A Balance](https://wayfarerlabs.ai/blog/gen-vs-rec)
|
| 64 |
+
- [Fast Audio Video World Models](https://wayfarerlabs.ai/blog/fast-audio-video-world-models-part-2)
|
| 65 |
+
- [Training Diffusable Autoencoders](https://wayfarerlabs.ai/blog/training-diffusable-autoencoders)
|
| 66 |
+
|
| 67 |
+
## Connect With Us
|
| 68 |
+
|
| 69 |
+
- 🎮 [Discord](https://discord.gg/dX4HW9Pt7Z) (1000+ researchers/developers) / model-specific & consumer facing [Discord](https://discord.gg/jT93EuasMA) (*Ask for a role!)
|
| 70 |
+
- 💻 [GitHub Organization](https://github.com/Wayfarer-Labs)
|
| 71 |
+
- 🌐 [Website](https://wayfarerlabs.ai/)
|
| 72 |
+
- 🧪 [Evaluation Platform](https://eval.wayfarerlabs.ai)
|
| 73 |
+
- 🐦 [Twitter/X](https://x.com/wayfarerlabs)
|
| 74 |
+
|
| 75 |
+
## Who Should Get Involved?
|
| 76 |
+
|
| 77 |
+
- **Researchers** - Collaborate on cutting-edge world model research
|
| 78 |
+
- **Developers** - Build applications with our models or contribute code
|
| 79 |
+
- **Roboticists** - Deploy models on embedded systems for real-time inference
|
| 80 |
+
- **Investors** - Learn about investment opportunities in open science AI
|
| 81 |
+
- **Partners** - Explore collaboration and integration possibilities
|
| 82 |
+
|
| 83 |
+
## License
|
| 84 |
+
|
| 85 |
+
This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.
|
| 86 |
+
|
| 87 |
---
|
| 88 |
|
| 89 |
+
*Built by the Overworld team in Providence, Rhode Island*
|